Use data from the same CERES instrument through the interval of interest.
This will minimize aliasing small instrument absolute calibration differences
(~ 0.5 to 1.0 W m -2) between CERES scanners into the record. The scanner
calibrations are not "tuned" into agreement. In this case, instrument
stability is much better than absolute accuracy. Estimated instrument stability
is to within 0.25 W m -2. Note that if regional trends are being examined, reducing
spatial sampling noise to less than 0.3 W m -2 requires averaging at least 9 monthly
mean 1 degree grid boxes for crosstrack data, and at least 36 grid boxes for
rotating azimuth plane data. The averaging can be in time and/or space as the
spatial sampling error is random and unbiased.
|
CERES Operations